Conclusion - camera shake is independent of sensor size.
This was precisely my point. And I think I've solved why we agree on this point, yet you said earlier that the crop factor should be used along with the focal length, with which I disagree.
It just so happens that the "standard" lens on a 35 mm camera often
has a focal length of 50 mm. By pure coincidence (let me emphasise
that - it is a coincidence ), on a 35 mm full frame camera, a 50
mm lens requires a shutter speed of 1/50, and a 100 mm lens
requires a shutter speed of 1/100. It just so happens, then, that
the shutter speed required is the reciprocal of the focal length.
By a happy accident.
I understand it's only an approximation, and trying to read too much into it falls apart because of the underlying assumptions. But it's still worth evaluating why we're finding different results. So let's delve into madnes for a moment
and use your example. Put a 50-mm lens on both a 35-mm frame camera and a camera with a crop factor of 1.6 (like the 20D/30D). The FoV on the 35mm camera is 46 degrees, while it's only 28 degrees on the cropped camera.
Now apply some camera "shake". Remember shake is not specifically a displacement, it's a motion, so assume 1 deg/s for your shake. If you used a shutter speed of 1 sec (for either camera), the camera would move 1 deg during the exposure. If you were shooting an object 10 feet away, this 1 degree of movement would cause about 2 inches of blur/ghost around the object. For an object 20 feet away, it would be about 4 inches, etc.
This sounds pretty crumby, so now apply the guideline for the 50-mm lens and use a shutter speed of 1/50. Now the motion will be only 1/50 deg, and the resulting object blur will be only 0.04 inch at 10 feet, 0.08 inch at 20 feet, etc. Clearly much better.
Note however that neither crop factor or FoV enter into any of this. Blur is the same regardless of the sensor size (as you concluded above), and you can thus disregard any crop factors when applying the guideline.
At this point I was happy until I asked myself what happens now with a 100-mm lens? (I wish I hadn't done that!) Focal length also didn't enter into any of the above blur considerations, so what's the reason for now going to 1/100 shutter speed if you follow the guideline? The only reason I can think of is this: when increasing focal lengths, we are generally shooting objects that are farther and farther away (whether it's simply because we can't get closer, or because we are unable to get the whole object in the frame and have to move back). As shown above, the blur increases linearly with distance from the camera. So to maintain an acceptable level of blur for objects that are farther away, you naturally want to increase shutter speed.
If, however, you continue shooting the same object at the same distance, then I think I now believe that focal length has no direct effect at all on blur. So if 1/50 was acceptable at 50mm, it will be similarly acceptable with 500mm, because the object is still the same distance from the camera, and the same camera shake and shutter speed produces the same acceptable blur. Objects farther away will be a mess of course.
This perhaps should have been obvious to me, but it explains why I can wander around shooting at 280, and have managed (in rare instances) to achieve okay shots at a 1/15---the rabbit was only 8 feet away. However shooting something at a couple hundred feet, and I want to see shutter speed at the focal length guideline or higher.
So my conclusions remains that crop factor would have no direct bearing in applying the guideline. My additional newfound conclusion is that the focal length technically doesn't matter either, but it's really the subject distance that matters. The focal length is used only as a loose convenience because it is usually linked to the subject distance.
The caveat on this, which has to cute effect of proving that we're all right
is that if you are shooting the same object with both cameras, but framing it both times to cover the same amount of the FoV, then the crop camera must shoot from a farther distance to make the subject fit in the fram (as it has a smaller FoV at the same focal length). To then achieve the same level of blur at this greater distance as the 35-mm camera, the shutter speed would need to be set higher. This is where I can see the crop factor being useful to combine with focal length in the rule. But it only arises due to the shooting constraint of framing the subject the same in both cameras.
Maybe this is the normal approach people are thinking of and why we were disagreeing. In my case I was thinking of the situation I described above, where the distance is the same in both cases, and you accept whatever FoV you get. This would be the case where you simply can't get closer, such as an airshow, autorace, wildlife, etc., and the subject generally won't fill the frame, even on the cropped camera. In this case, the effect of blur is the same for both cameras, regardless of the crop factor, and you can simply use focal length in the guideline. However, as mentioned above, it's not even the focal length that really matters so much as the subject distance. In any case, same distance means same shutter speed would apply for both cams at the same actual focal length, regardless of crop factor.
Well this exercise was certainly educational for me! It's nice that I have a better understanding of the focal length guideline, and how it relates (or doesn't) to different shooting situations. I'm hoping someone else got something out of it, too.
The end result for me? Right where I started---crank the shutter speed as high as I can get it with the aperture I want and the ISO I can tolerate
--
Regards,
K
Ontario, Canada
(See my profile for equipment list.)
